Changes in sibling configurations for American preschool children.

This article was downloaded by: [University of Sussex Library] On: 18 January 2015, At: 02:54 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Biodemography and Social Biology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/hsbi20

Changes in Sibling configurations for American preschool children David J. Eggebeen

a

a

Department of Human Development and Family Studies and Population Research Institute , The Pennsylvania State University , University Park, Pennsylvania Published online: 23 Aug 2010.

To cite this article: David J. Eggebeen (1992) Changes in Sibling configurations for American preschool children, Biodemography and Social Biology, 39:1-2, 27-44, DOI: 10.1080/19485565.1992.9988802 To link to this article: http://dx.doi.org/10.1080/19485565.1992.9988802

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

Downloaded by [University of Sussex Library] at 02:54 18 January 2015

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-andconditions

Changes in Sibling Configurations for American Preschool Children

David J. Eggebeen


Department of Human Development and Family Studies and Population Research Institute The Pennsylvania State University University Park, Pennsylvania ABSTRACT: This paper uses data drawn from the 1940 through 1980 Public Use Microdata Samples of the U.S. Census of Population to document sibling configurations from the child's perspective. Changes in four aspects of siblings are examined for five cohorts of white and black preschool-aged children: number, birth order distributions, spacing intervals, and sex composition. Changes in fertility behavior of adults in the post-war era had a profound effect on the structure of sibling systems experienced by children. Successive cohorts of preschool children show a rise in number of siblings through the early post-war years before showing sharp declines in number of siblings through the 1960's and 1970's. These shifts in size of sibling sets are reflected in changes in the proportion of each cohorts who are first born and only children, both of which have increased substantially by the 1980 cohort. The 1940 and 1980 cohorts have similar proportions of children with short intervals. However, the middle cohorts show the effects of the quickened pace of fertility with substantial proportions of children with comparatively short birth intervals. Finally, substantial shifts across cohorts in several measures of sex composition of children are observed. Most significantly, there is a marked decline in the proportion of children experiencing an opposite-sex older sibling.

Since the Great Depression, adults' childbearing behavior has dramatically changed. The total fertility rate increased from 2.30 in 1940 to 3.77 by 1957, before falling to 1.84 in 1980 (National Center for Health Statistics, 1988). Furthermore, these fertility shifts took place in almost every subgroup in American society, although some, such as blacks and Catholics, were more involved than others (Rindfuss and Sweet, 1977). The causes and consequences of these changes have preoccupied many American demographers for some time. How these shifts have affected the lives of children, however, has received surprisingly little attention. One major consequence of fertility changes is expressed in the experience of siblings. As women have changed the timing and number of chil-

dren they bore over the past forty years, the sibling experience of successive cohorts of children must inevitably have changed.1 This paper represents an attempt to document these shifts. Data drawn from the most recent censuses will allow us to see how preschool-aged children from successive birth cohorts systematically differ in sibling configurations. The focus will be on documenting shifts in four components of the sibling system: number of siblings, ordinal position, spacing intervals, and\ sex composition. A basic premise of 1 Not all changes in the number of siblings over time (from the child's perspective) are the result of shifts in childbearing. Changes in the adoption rates or the likelihood of more distantly related children living in the home can also affect the experience of siblings.

27


28

Eggebeen

this paper is that sibling configurations matter. That is, number of siblings a child has, his/her birth order, the intervals between children, etc., have a significant effect on that child's development. It becomes important, then, to track the changing structure of sibling systems across cohorts of children. I begin with a brief review of the four components of siblings to be examined, followed by a discussion of the data used to derive the population estimates, the findings, and discussion.

COMPONENTS OF THE SIBLING SYSTEM Of all the configuration variables, number of siblings has probably received the most research attention. A number of studies have explored the effects of size upon the "internal" workings of families. Some report that behavioral control, methods of rule implementation, and affection patterns in families are affected by the number of children (Elder and Bowerman, 1963; Kidwell, 1981, 1982; Nye et al., 1970; Peterson and Kunz, 1975; Scheck and Emerick, 1976). Explanations for these findings center upon the theoretical importance of numbers on the interactions of members of a small group. Others have focused on the consequences of various family sizes for children's mental ability, school achievement, adult careers, or other outcomes. Some studies suggest that children from small families score higher on intelligence tests and do better in school than children from large families (see Rodgers and Thompson, 1985/1986, for a review). The conflu-

Social Biology

ence theory, for example, suggests that children's intellectual development is affected by two components, the intellectual level of the environment in which they are raised, which is affected by family size, and the intellectually facilitating influence of having a younger sibling to teach (Zajonc, 1976; Zajonc et al., 1979). Judith Blake (1981, 1989) suggests a "dilution model" of parental inputs as a variation on the confluence model. Here, simply the more children, the more parental resources are divided and hence "the lower the quality of the output" (1981, p. 422). She maintains that even when major cultural, period, and socioeconomic variables are taken into account, increasing family size has an important negative influence on children's educational attainment. Despite differences in these two theories on the role of birth order, both theories agree that having a large number of siblings is a disadvantage for a child which even high family income and high parental education does not easily offset. What can we expect in the way of changes in number of siblings across cohorts of preschoolers? Since the baby boom did not represent a return to the large families, but rather a movement away from spinsterhood, childless marriage, and toward the one-child family, one might expect relatively small changes in average completed family sizes of children in the 1940 through 1980 decades (Preston, 1976; Westoff, 1978). However, the baby boom was also characterized by a shift in the timing of births with more women having their children bunched together at younger ages (Westoff, 1978). Other things being equal, this


Vol. 39, No. 1-2

Sibling Configurations

bunching of births will tend to increase family size as experienced by preschoolers in the 1950's compared to the early decades and the most recent decades. Thus we might still expect to see some increase in the average number of siblings. Conversely, a widening of the birth spacing in the baby bust years could function to accentuate the decline experienced in the preschool years from 1960 onward. A second dimension of sibling structures to be examined is birth order. Although ordinal position has attracted almost as much attention through the years as family size, it is much less theoretically developed and much more controversial. As Schvaneveldt and Ihinger (1979) note, studies in this area "are not known for innovative methodology or carefully formulated research designs. For the most part, this research has been ex post facto, disjointed, tangential, and motivated by curiosity rather than theory" (1979, p. 454). Some researchers have argued that this line research is bankrupt and should be abandoned (Schooler, 1972). Others see merit in continued investigation, although they are critical of much of the research (Adams, 1972; Blake, 1989; Rodgers and Thompson, 1985/1986). Much of this work concentrates on birth order as a determinant of educational achievement, intelligence, aspirations, or various aspects of personality, and it emphasizes differences between first borns and later borns. Although a number of different theoretical orientations have been employed to explain differences by birth order (see Rodgers and Thompson, 1985/ 1986), it can be argued that an underlying commonality to a number of

29

them is the emphasis on the uniqueness of being first born. With the birth of the first child, parents begin to "try out" their theories of child raising. The birth of the first child also initiates interactions between parents and other social institutions and groups such as schools, church, peer groups, and community groups, which, to a greater or lesser extent, affect the family system. Thus, with the birth of the first child, and with each subsequent birth, the family system becomes more "open." The oldest child, then, acts as a trend setter or trail blazer in this process. The feedback from these first experiences is likely to change parenting behavior for subsequent children. Later-born children have parents who are more experienced and thus more likely to be relaxed, not tentative, and possibly more tolerant or permissive than they were for the firstborn child (Bossard and Boll, 1960; Schvaneveldt and Ihinger, 1979). This line of reasoning fits well with some commonly found birth-order patterns. For instance, first-born children have been found to elicit higher parental expectations for behavior and performance, get more parental time and attention, to be more conservative, have lower levels of sexual activity in adolescence, to have higher selfesteem, and have higher levels of educational attainment than later-born children (Blake, 1989; Kammeyer, 1966; Kidwell, 1982; Lasko, 1954; Lewis and Kreitzberg, 1979; Reiss, 1967; Rodgers and Rowe, 1988). The distribution of cohort members by birth order depends, in part, on the sizes of families. Where the family sizes of children are large, the ratio of first born to later borns will be small.


30

Eggebeen

As the average family sizes of children decline, the proportion of children who are first born will increase. We can anticipate, then a relatively small proportion of first borns for the baby-boom cohorts compared to the preceding and following cohorts of children. A third dimension of sibling systems is birth intervals. Demographers have been interested in birth intervals because of their centrality in understanding the biological and social dimensions of the family formation process (Hastings and Robinson, 1975; Menken, 1975; Tsui, 1982). Short intervals imply more rapid childbearing, which may result in an earlier age at last birth or in larger families (Spanier et al., 1985). Longer birth intervals may have been intended delays because of labor force involvement, residential or marital status changes, as well as unintended delays because of illness or other involuntary factors (Groat, 1982; Jones, 1981; Potter, 1963). Social scientists also have been interested in the spacing of children because of its potential effects on child development. Close spacing of children has been linked consistently with negative effects on cognitive development. In contrast, wide spacing has been found to be beneficial for the intellectual development of both the younger and older child (see Wagner et al., 1979, for a review of research in this area). There also has been some work documenting differing parenting styles by the spacing of children. Parents of adolescents that are spaced less than two years apart use more physical force to control their offspring, although this does not lead to more parent-child conflicts (Edwards and Braunburger,

Social Biology

1973). Kidwell (1981) also reports that for adolescents, average sibling spacing is negatively related to perceived parental punitiveness and positively related to perceived parental reasonableness and supportiveness. Given what we know about the changing amount and pace of fertility since 1940, we anticipate relatively large intervals for preschool children in families with at least two children in the pre-baby-boom cohorts. These intervals should decline as the pace of childbearing increased in the 1950's. The most recent cohorts, however, should show a return to longer intervals, reflecting the "baby bust" fertility patterns (U.S. Bureau of the Census, 1978). The final structural dimension examined is that of sex of siblings. It is commonly asserted that the sex composition of the children in a family has important consequences (Schooler, 1972; Schvaneveldt and Ihinger, 1979). Nevertheless, there has been relatively little done theoretically or empirically to substantiate these claims. One exception to this is the crosssex assimilation hypothesis developed by Orville Brim (1958). Based on the work of George Herbert Mead, Brim hypothesizes that children with siblings of the opposite sex will display more traits of the opposite sex than will children who have siblings of the same sex. Also, he suggests that children having older opposite-sex siblings will display the highest number of the traits of the opposite sex. Support for this hypothesis is found not only by Brim, but also by Kammeyer (1967). Extensions of Brim's cross-sex assimilation hypothesis can also be found in the work of Reiss (1967) and Rodgers (1983) ex-


Vol. 39, No. 1-2


plaining adolescent sexual behavior and attitudes. Both found that sex composition, especially that of the older siblings relative to the child, has a significant effect on a child's attitudes about premarital sex or their intercourse behavior. In addition to Brim's model, several researchers have suggested that a "unique gender status" in sex composition can account for some differences between children. That is, being the only male or only female in a family is a situation which can have special consequences (Elder and Bowerman, 1963; Schvaneveldt and Ihinger, 1979). Kidwell (1982) tests several hypotheses derived from this notion in her examination of determinants of self-esteem in adolescent boys. She finds that the disadvantage of a middle-born birth order is partially offset by characteristics which enhance the distinctiveness of middle-borns, prominent among them the uniqueness associated with being the only male among a number of female siblings. MATERIALS AND METHODS Data for this paper are drawn from the 1940 through 1980 Public Use Microdata Samples of the U.S. Census of Population. A children's file for each census year was created, providing a representative sample of all children living in households with their parents). The record for each child contains information on the child, his or her parent(s), and all sibling(s) of that child living in the home. By definition, children living in institutions or in households where they are not related to the head of the primary family or subfamily were deleted from the files.

31

It is important to note that in this paper children are the unit of analysis. Most previous demographic research has taken the perspective of the household, family, or adults. As demonstrated by Preston (1976), however, findings from adult-based data are not necessarily indicative of the experience of children. If we wish to know what the consequences of changing fertility are for the sibling experience of children, we must use children as the unit of analysis. The analysis is restricted to white and black children aged zero to five at the time of the decennial census. Thus, information is available for five sixyear birth cohorts of children, starting with the 1934-39 birth cohort and ending with the 1975-79 birth cohort. The reason only preschoolers are examined and not children of all ages is because of an important limitation in these data, that is, there is no information on children who are 18 or older. This problem is minimized by limiting the analysis to preschool-aged children. Nevertheless, the reader should be aware that an unknown (but probably small) proportion of preschool children who are designated here as having no older siblings actually have a sibling who is 18 or older.2 Siblings are defined as any "own child" of the head of the primary, sub-, 2 Of course, by limiting the analysis to preschool children, we miss younger siblings not yet born. Thus, the actual number of siblings experienced by these cohorts as they age through their life course will be somewhat larger than the estimates derived from these data. The number of additional siblings born undoubtedly varies across cohorts with changes in number of children born and fluctuations in spacing intervals. Distributions of siblings could be derived from completed fertility data for women. Yet, this method is also not without bias as well, especially from selective mortality.


32

Eggebeen

or secondary family in the household in the same primary, sub-, or secondary family as the preschool child. No distinction is drawn in this paper between half, step, or whole siblings. Undoubtedly there are important race and cross-time differences in the likelihood of having certain types of siblings. Unfortunately, the marital history information that was collected is not consistent across the five censuses. Since this information greatly complicated the definition of types of siblings, this research was not pursued. RESULTS NUMBER OF SIBLINGS

In Table 1 we see that the post-war fertility fluctuations caused large cohort differences in the distributions of children by number of siblings. Looking first at whites, we see that in the baby boom years, there were increasing proportions of preschoolers living in families with three or four siblings. As one might expect, this was accompanied by a decline in the proportion of children with no siblings. However, there was also a decline in the proportion of children living in families with six or more siblings. Thirteen per cent of all preschoolers were living in families this large in 1940; but by 1960 the proportion of these preschoolers was nearly half that of the 1940 cohort, 7.6 per cent. In fact, the proportion of children living in a family with six or more siblings declined steadily across all five decades, and by 1980 less than 2 per cent of white preschool-aged children were living in this family context. In more recent years, we see the consequences of the well-known fertil-

Social Biology

ity declines and the rise of the twochild family. From 1960 to 1980, the proportion of white preschoolers without siblings nearly doubles (from 14 to 26 per cent) while the proportion of children living with just one other sibling jumps from 30 to 42 per cent. Thus, nearly 70 per cent of all preschool children in 1980 lived with just one sibling or were only children compared to only 44 per cent in 1960. The patterns for blacks are somewhat different, reflecting the higher average number of siblings experienced by black children in their preschool years. For instance, the proportion of black preschoolers living in families with a large number of siblings (6 +) is consistently three times that of whites across all decades, while the proportion living in families with one or two children varies from less than half that of whites in the 1940's and 1950's to about four-fifths that of whites in 1980. BIRTH ORDER

The bottom panel of Table 1 displays the distribution of preschoolers by birth order.3 As expected, the increasing family sizes of children soon after World War II are reflected in a decline in the proportion of these children who are first born between the first and second cohorts, while the proportion of children of higher order 3 Children who are the products of multiple births are all assigned the same birth order. For example, twins who are the third and fourth children in a family are both designated as being third-borns. One result is that I overestimate the proportion of "true" first-born children. Fortunately, first-born multiple births are rare, occurring at a rate of 6.7 per 1,000 live births for whites in 1964 and 7.8 per 1,000 per blacks (Heuser, 1967).


i p

TABLE 1 PERCENTAGE DISTRIBUTIONS OF PRESCHOOL CHILDREN LIVING WITH AT LEAST O N E PARENT BY NUMBER OF SIBLINGS LESS THAN 18 AND BIRTH ORDER No. op SIBLINGS AND

BIRTH ORDKR

1940

White

Number of siblings 1 2 3 4 5 6+ Birth order 1 2 3 4 5+ N

1950

1961)

1970

1981)

"

Black

White

Black

White

less than 18 23.1 9.8 29.1 15.2 16.8 18.0 10.7 12.1 7.2 12.2 13.3 32.6

15.2 31.4 21.6 12.7 7.4 11.8

7.8 13.6 15.7 15.8 11.8 35.4

14.4 29.5 26.1 15.4

34.9 25.3 14.2 9.4 16.2

28.9 29.1 17.8 10.1 14.1

17.3 17.4 15.4 13.6 36.3

30.2 28.8 20.4 10.9

7.1 7.6

Black

White

10.3 17.2 16.7 16.8 12.1 26.9

19.0 33.4 23.3 13.1

21.8 20.2 17.4 14.2 26.4 2,884

34.0 29.1 18.6 10.0

5.9 5.2

Black

White

Black

,14.6 22.0 17.8 16.9 11.9 16.8

26.2 41.8 21.0

26.2 22.9 16.9 13.9 20.1 2,721

42.4 34.6 15.6

34.2 31.2 17.5

5.1 2.3

9.5 8.5

7.5 2.4 1.2

21.2 33.8 20.7 13.0

era

6.7 4.6

C os

a

5

era

11,541

19.0 18.4 14.1 12.1 36.4 1,439

25,024

3,682

SOURCE: 1940. 1950. 1960. 1970. and 19S0 PUMS files of the U.S. Census of Population.

9.7

20,842

8.3

17,162

14,814

2,657

i

1


34

Eggebeen

Social Biology

births increased. The ensuing cohorts, ond child comes along. This change in however, have increasingly larger pro- timing, then, would accentuate the portions of first borns. For whites, 29 proportion of children who are singleper cent of the 1950 cohort were first- tons for at least the first five years of born children. By comparison, 42 per their life. cent of the 1980 cohort are first-born, The results in Table 2 seem to supan increase of nearly 45 per cent. For port the later notion. Here we see that black preschoolers, we find that the the average interval in years between a proportion first-born nearly doubles, first-born and the second-born (only from 17 per cent in the 1950 cohort to for those with at least one sibling at the 34 per cent in 1980. Thus, the propor- time of the census) has increased from tion of black preschoolers who are 1.9 years in 1960 to 2.3 years in 1980 for first-born appears to be converging whites. This suggests that we have to with that of whites. be cautious about reading too much An important distinction is drawn into the 82 per cent increase in the proby some researchers in the birth order portion of only children among preliterature between first-borns and only- schoolers from 1960 to 1980. born children. Blake (1981) notes that Is there any evidence that singletons while being an only child is widely re- may be on the rise? Table 3 displays garded by Americans as a significant handicap, there is almost no empirical TABLE 2 support for such a stereotype. In fact, MEAN INTERVAL BETWEEN THE CHILD AND THE in her own analysis, she finds that only NEXT YOUNGEST SIBLING FOR FIRST-BORN children are, on average, intellectually CHILDREN WITH AT LEAST O N E SIBLING IN YEARS superior and achieve higher educa- MEAN INTERVAL 1940 1950 1960 1970 1980 tional and occupational status. We can White see in the top panel of Table 1 that the Mean 2.06 1.97 1.91 2.14 2.30 proportion of preschoolers who are s.d 1.11 1.05 0.97 1.03 1.09 N 1,361 3,431 3,308 2,582 2,400 only children increased from a low of 14.4 per cent in 1960 to over 26 per cent Black Mean 1.86 1.64 1.61 1.78 2.13 of all preschoolers in 1980. As a cons.d 0.89 1.00 0.94 1.05 1.28 sequence, the proportion of first-borns N 133 349 333 317 346 who are only children increased from SOURCE: Same as Table 1, 48 per cent to 62 per cent in 1980. There are two possible reasons for this observed increase. First, we are seeing TABLE 3 the consequences of adults choosing to PROPORTION OF FIVE-YEAR-OLD CHILDREN W H O have only one child. However, there ARE THE ONLY CHILD LIVING IN THE HOUSEHOLD may be another possibility. Given that AT TIME OF CENSUS we are limiting our focus to children Race 1940 1950 I960 1970 1980 aged zero to five, we may be seeing the White.... 14.4 7.6 7.6 7.9 14.6 results of an increase in birth intervals. (1,981) (3,830) (3,386) (3,223) (2,383) That is, across cohorts, first-born chil- Black.... 8.6 4.4 7.9 8.0 16.4 (257) (527) (469) (511) (445) dren may be spending longer periods of time as the only child before the secSOURCE: Same as Table 1.


Vol. 39, No. 1-2


the proportion of five-year-olds who are only children. Here we see a large jump in the proportion of only children between 1970 and 1980 for both black and white children. Certainly some of these children will eventually acquire siblings. However, given these children have already gone nearly six years without gaining a sibling, which is more than two-and-a-half times the average birth interval of both races, one can reasonably assume that the majority of these children will continue their childhood as singletons. Even those who acquire a sibling will have spent their early critical years as only children—a socialization experience strikingly different from that of those children who either are born into an existing sibling system, or in their preschool years acquire siblings to whom they must adjust. BIRTH SPACING

I begin with an examination of birth intervals to the next youngest sibling.4 4 The terms birth interval and birth spacing will be used interchangeably. The analysis focuses on the age difference between living siblings and does not take into account the effects of mortality. Our measures of birth intervals are somewhat crude because the age of individuals is only recorded as their age in years as of April 1, the year of the census. This, of course, introduces some error when calculating birth intervals. For example, a child who is estimated to be three years older than a sibling may actually differ in age by two years plus one day or by four years minus one day, depending on the month and day of birth of a child and the month and day of birth of their sibling. However, population estimates of intervals from a particular census are unlikely to be biased if one assumes that the distribution of children by time of year born is approximately the same as the distribution of siblings by time of year born. Likewise, comparisons across censuses will be unbiased if any differences in the distributions between children and their siblings in time of year born remain constant. An additional source of error in estimating intervals comes from the fact that some children will be the same age as one or more of their siblings be-

35

In order to minimize the confounding effects of age and interval, this part of the analysis is restricted to five-yearolds with at least one younger sibling. As is evident in Table 4, the likelihood of having a younger or older sibling varied considerably across cohorts, and this finding needs to be kept in mind when reviewing the changes in intervals. Examining the changes for whites first, we observe a marked increase between 1950 and 1960 in the proportion of children having a sibling spaced within two years. Most of this gain came from a decline in the proportion of children with siblings of a moderately close age (two to four years), while the proportion in the remaining categories remained relatively stable. The changes in the post baby-boom years presents a different pattern. There was a dramatic decline in the proportion of five-year-olds with short intervals (less than two years); in 1980 it is half that of 1960. However, most of this decline is accounted for by the increase in the proportion of children with wide intervals. The proportion of children with a sibling four or more years younger increased from 21.5 per cent in 1960 to 29.8 per cent in 1980. The pattern of change for blacks is similar to that of whites, except for the greater likelihood of blacks having shorter intervals in every cohort. Note, however, that this race difference appears to have declined since 1950. cause of multiple births, blended families from remarriage, or possibly from adoption. The consequences of spacing could potentially be quite different depending on whether one's own sibling of the same age is a twin, an adopted child, or the result of remarriage. Unfortunately, with census data it is not possible to distinguish between these various reasons for multiple same-age children within a given family.

Eggebeen

36

Social Biology

TABLE 4 DISTRIBUTION OF FIVE-YEAR-OLDS BY INTERVAL TO NEXT YOUNGEST SIBLING


Interval in Years

White < 1.0 1.0-1.9 2.0-2.9 3.0-3.9 4.0-4.9 5.0-5.9 N % with no younger sibling Black < 1.0 1.0-1.9 2.0-2.9 3.0-3.9 4.0-4.9 5.0-5.9 N % with no younger sibling

1940

1950

1960

1970

1980

0.3 18.7 36.4 24.4 13.0 7.3 1,056 46.0

0.2 18.5 36.3 22.5 14.4 8.2 2,289 39.3

3.2 24.2 30.2 20.9 13.0 8.5 2,069 38.9

2.6 22.1 32.3 19.9 14.3 8.7 1,660 48.4

2.3 11.5 30.5 26.2 18.8 11.0 1,109 52.6

0.0 26.3 43.6 20.1 7.3 2.8 179 29.0

0.3 32.3 41.6 13.0 7.3 5.5 339 26.3

4.9 35.5 31.8 16.1 7.6 4.2 330 28.6

7.1 32.8 27.0 17.2 10.5 5.4 296 41.6

3.8 24.6 25.1 17.5 13.7 15.2 211 52.6

'

SOURCE: Same as Table I.

Table 5 presents the distribution of all preschool children by interval to the next oldest sibling. For whites, the changes from 1940 to 1960 show a marked increase in the proportion of preschoolers with a very short interval (less than two years) between them and their next oldest sibling. There are either no changes or declines for the rest of the intervals. We observed a reversal of these trends from 1960 to 1980. The proportion of white preschoolers with very short intervals nearly halves, from 25 per cent to 13 per cent, while the proportion whose next oldest sibling is six years or more increased from 15 per cent to nearly one in every five preschoolers. The changes for blacks follow a different pattern. In 1940 nearly 23 per cent of preschoolers had an older sibling who was less than two years older.

By 1960, however, the proportion of children in this situation had increased by two-thirds to 37 per cent. In contrast to whites, most of this shift came from declines among the moderate birth intervals (two to three years), which characterized 56 per cent of all black preschool children with an older sibling in 1940, but only 44 per cent by 1960. There was very little change in the proportion of children with wide spacing. Like whites, the changes from the baby boom years to the present are marked by an increase in spacing. Evaluating the significance of the shifts in birth spacing between a child and their siblings in isolation from the changes in the other components of the sibling structure is problematic. However, we can examine the simultaneous consequences of changes in family size and birth spacing through measures of family density.

Vol. 39, No. 1-2

37


TABLE 5 DISTRIBUTION OF PRESCHOOL CHILDREN BY INTERVAL TO NEXT OLDEST SIBLING


Interval in Years

White < 1.0 1 0-1 9 2 0-2 9 30-3 9 40-4.9 5.0-5.9 > -6 0 A?

% with no older sibling Black < l0 1.0-1.9 2 0-2 9 3 0-3 9 4 0-4 9 5.0-5.9 > -6.0 N % with no older sibling

1940

1950

I960

1970

1980

2.2 13.8 28.8 18.7 11.4 7.0 18.0 7,537

2.7 18.1 27.0 18.7 11.4 6.8 15.2 17,869

2.8 21.9 27.7 16.8 9.8 6.7 14.5 14,648

1.8 18.0 26.4 18.4 11.8 7.5 16.0 11,354

1.9 11.3 24.1 20.6 13.7 8.8 20.2 8,599

34.6

28.5

29.7

33.8

41.2

2.7 20.1 36.7 19.6 7.6 3.2 10.7 1,168

3.9 27.7 33.1 14.5 6.6 3.4 10.7 3,020

5.1 31.9 28.7 15.0 7.0 3.5 8.8 2,256

5.4 27.1 24.5 14.6 8.6 5.7 13.9 2,004

3.6 16.0 19.8 14.6 11.8 8.6 25.6 1,768

18.8

17.1

21.1

25.8

33.5

SOURCE: Same as Table I.

A convenient measure of density which combines these two aspects of sibling structure in a theoretically meaningful way was developed by Kidwell (1981). The formula for this measure is:

y where: D = sibling density, and 5,- = the number of years between respondent (5) and each of his/her siblings (/). Summing the reciprocals of the spacing between the respondent and each of his/her siblings gives decreasing weight to siblings who are spaced further away. For example, two children may each have one sibling, but if one has a sibling three years different in age while the other child's sibling is one year away, the former child

will score lower on this density measure. The assumption here is that the particular combination of both the number and the "distance" of a child's siblings are important factors in determining the impact these siblings have on the child (Kidwell, 1981). Figures 1 and 2 contain the distributions of five-year-olds by density. Limiting the analysis to five-year-olds was done to include younger siblings. The major shifts we observe for whites (Figure 1) is the increased proportion of children with very low densities from 1960 onward. For instance, over half of all white children have a density between zero and 0.5 in 1980, up from one-third of all children from 1940 to 1960. Black five-year-olds reside in markedly more dense families than whites in every decade, although the

Eggebeen


38

trend since 1960 is that of declining differences (Figure 2). As was true for whites, 1960 marked the peak time in terms of family density. After this date, large declines in number of siblings and increasing birth intervals meant the average density in 1980 was half that of 1960. Also, like whites, the changes in the distribution can be characterized by an increasing concentration of five-year-olds in one or two categories of density. In 1940, one-third of black children had a family density between zero and 1.0; this per cent increased to nearly 60 per cent by 1980. Thus, both white and black five-yearolds are far less likely in 1980 to live in crowded families than has been the case for any time since 1940. To summarize, calculations of the age differences between preschool children and their siblings reveal large changes across cohorts. The average interval between a child and their next youngest or next oldest sibling declined during the baby boom cohorts, but then increased with each subsequent cohort. Changes in a measure of family density which reflects both the number and the spacing of all siblings in the family revealed similar patterns of change for both whites and blacks. SEX COMPOSITION

From the work of Brim (1958) discussed above, we know that having an older sibling of the opposite sex represents a special milieu. Table 6 contains the changes in sex composition of older siblings. For whites, we see small changes over time in the proportion of males or females with only oppositesex older siblings. However, the proportion of children with older siblings of both sexes has declined sharply

Social Biology

across the last three cohorts. Hence, since 1960, there has been an overall decline in the proportion of white preschoolers with at least one opposite sex older sibling, from nearly half of all children in 1960 to just over a third by 1980. Black children present a different pattern from that of whites, although the overall differences in sex composition appear to be diminishing. The proportion of black males and females with only opposite-sex older siblings appears to increase consistently over time, from about 13 per cent in 1940 to 25 per cent for females and 23 per cent for males. This increase is offset, however, by the comparatively large declines in proportions of children with older siblings of both sexes. As a consequence, the proportion of black children with at least one older sibling of the opposite sex is consistently higher in each decade than whites, although this difference has declined over time. Some researchers have suggested that being unique in terms of gender among one's siblings can have special consequences for a child (Elder and Bowerman, 1963; Kidwell, 1982; Schvaneveldt and Ihinger, 1979). Of course, we cannot determine who among our preschoolers will spend their entire childhood as the only male or female child; but we can look at changes across cohorts in the proportion of children who at age five are still the only male or female among their siblings. Looking first at whites (Table 6), we see that the proportion who are the only males or only females in their family has increased since 1950. Only males, for instance, increased from 21.3 per cent of all five-year-old males in 1950 to 29 per cent in 1980. Only

Vol. 39, No. 1-2


100%

75% -


50%

25% -

^ H >0.0-0.5

H I >1.5-2.0

•

>2.0-2.5

FIG. 1.—The distribution of five-year-old white children by sibling density.

100%

75% -

50% -

25%

>0.0-0.5

CH M.5-2.0

•

>2.0-2.5

FIG. 2.—The distribution of five-year-old black children by sibling density.

39

40

Eggebeen

Social Biology

TABLE 6 GENDER STATUS OF FIVE-YEAR-OLDS 1940

1950

1960

1970

1980

White Males No siblings The only male Male & female sibs. N

14.1 21.9 64.1 1,024

7.7 21.3 71.0 1,999

6.8 23.5 69.7 1,762

7.2 24.0 68.7 1,643

14.9 29.0 56.2 1,239

Females No siblings The only female.... Male & female sibs. N

14.9 23.1 62.1 954

7.6 23.3 69.2 1,829

8.4 24.1 67.6 1,624

8.6 26.8 64.6 1,577

14.4 32.1 53.5 1,143

Black Males No siblings The only male Male & female sibs. N

9.2 12.2 78.6 131

4.2 11.4 84.4 263

9.7 16.8 73.5 226

7.6 18.3 74.1 263

16.9 26.2 57.0 237

Females No siblings The only male Male & female sibs. N

7.9 12.7 79.4 126

4.6 15.2 80.3 264

6.4 14.0 79.7 236

8.6 20.9 70.5 244

15.9 24.5 59.6 208


Gender

SOURCE: Same as Table 1.

females show a similar increase over this time period, from 23 per cent to 32 per cent. We see approximately the same pattern of changes for blacks, with the proportion of unique gender children increasing across the most recent cohorts. Obviously, as the number of siblings declines, the likelihood of having a unique gender status relative to ones' siblings increases. However, the significance of this may not be as great as it might first appear. It is possible that the salience of this factor may be minimal where a child is the only male in a two-child family. Here its effects may be overshadowed by the fact that the child either is a first-born or a lastborn—sibling roles which themselves set a child apart. The unique gender ef-

fect suggested by past research (the mere fact of being the only boy or only girl works to set one apart from one's siblings in terms of parental attention and expectations) may operate most powerfully for children with moderate to large sibling sets. Given the large increases in two-child families and the decline in large families in the three most recent cohorts, it is possible that the recent increases in the proportion of unique gender-status children may not mean as marked a shift in developmental context as first appears. DISCUSSION Clearly, the sibling experience of preschool-aged children as characterized by number of children and birth


Vol. 39, No. 1-2


order has gone through considerable change across these five cohorts. Despite this change, little thought has been given to the consequences of these cohort shifts. If the unique socialization experience of first-borns sets them apart from later-borns on some characteristics, what are the consequences for these cohorts of being increasingly made up of first-borns since 1960? The same question can be asked in regard to the changes in family size. Judith Blake (1985), for one, suggests that the declines in family size were a major factor in the rise in levels of education for men in the last half century. She concludes from her analysis that the increased upward educational mobility that Featherman and Hauser (1975) found was the case for men was actually due almost entirely to the experience of men from small- and medium-size families. Thus, as Hernandez (1986) notes, Blake's research highlights the fundamental role number of siblings plays in the status attainment process by defining the conditions under which social mobility has increased in America. To the extent that birth order and number of siblings are salient features of childhood, and they influence in patterned ways the kinds of adults that individuals become, we must take them seriously as mechanisms of social change in society when the distribution of cohorts of individuals by these characteristics change. Perhaps we could learn something about the consequences for the 1980 cohort of having high proportions of children who are first-born, only children, and small sibling sets by studying the 1940 cohort, which had

41

similarly high proportions of these characteristics. What might be the consequences of the declining likelihood of preschool children having closely spaced siblings or living in crowded families? One might anticipate that the negative features of short intervals documented in the literature, e.g., poorer health of children, reduced individualized parental attention, less sibling competition, more positive parenting styles, poorer cognitive development, will be less common among the more recent cohorts of children relative to the past. Of course, a test of this hypothesis is not possible with census data. The social, psychological, and developmental consequences of the observed cohort changes in sex composition for the lives of children remain unexamined. Will successive cohorts of boys have fewer "feminine" traits as fewer and fewer of them are exposed to older sisters? Existing research suggests that this will be the case. But on the other hand, the social world of preschool children of the recent cohorts is increasingly marked by the experience of nonfamily child care arrangements. Preschool children with mothers in the labor force increased from 29 per cent in 1970 to 43 per cent by 1980 (Eggebeen 1986,1988), and the figure is projected to climb to over 70 per cent by 1995 (Hofferth and Phillips, 1987). In the majority of cases, the mothers of these children work full-time (Eggebeen 1986), suggesting a growing proportion of children are likely spending a significant part of their day in childcare. These experiences with nonfamily childcare arrangements mean children are probably much less "closed


42

Eggebeen

off from nonfamily individuals than has been the case in the past. Thus, the consequences of the decline of siblings with its loss of opportunity for crosssex modeling, may be offset to some degree by the gain in opportunities for cross-sex modeling and imitation that exists for children in day-care settings. On the other hand, there have been changes which might serve to offset to some degree these potentially positive gains. Increasingly, preschool children are spending time in some form of group care outside the home (Hoffreth and Phillips, 1987). Of course, the amount of time this involves varies widely. For some children this means a nursery-school setting for a small part of their day a year or two before they start kindergarten. At the other extreme are children whose parents are working full-time, and group care is a major part of their lives and in some cases has been since birth—a situation which is becoming more common. For example, the proportion of children whose mothers are employed full-time and who are in group care has steadily increased from 6 per cent of all children under age three in 1965 to 16 per cent in 1982 (Hoffreth and Phillips, 1987). This growing trend raises questions about its consequences. How does this group care experience differ from growing up in a home with a large number of closely spaced children? Do children compete in analogous ways for the attention and affections of the teacher/childworker? These are questions that cannot be answered with census data, but they merit attention in the future. Another social change which may have important implications for understanding the consequences of changing

Social Biology

sibling structures is the recent rise in mother-only families due to divorce and nonmarital childbearing. One implication is that the potentially positive gains to children of declining number of siblings is at least partially offset by declines in the number of parents living with children (Hernandez, 1986). Furthermore, the changes in gender composition of families of children is affected not only by the gender composition of the children, but also by the fact that children of divorce usually reside with their mothers (U.S. Bureau of the Census, 1988). Finally, divorce and its aftermath—custody, visitation, and eventually remarriagepresent the possibility that children's experience of siblings may be much more complex and dynamic today than was the case for earlier cohorts. Thus, we need to be cautious in assessing the consequences of any single change in isolation from other social changes that are taking place in the lives of children. Finally, most of this discussion has focused on the short-term implications of shifts in sibling structures, but clearly there are long-term consequences as well. As many others have noted, declines in numbers of children per family has significant implications for parent care in later life (Bengtson and Dannefer, 1987; Eggebeen, 1991; Hagestad, 1986; Preston, 1984). However, the long-term implications, if any, of changes in spacing, birth-order distributions, and gender composition are less clear. In conclusion, the consequences of the declines in family sizes since the late 1960's have been given considerable attention. Virtually ignored, however, has been its considerable impact on the life course of children. Most

Vol. 39, No. 1-2


children today still grow up in families with parents and siblings, but many aspects of life in families have changed profoundly. We are only beginning to examine the consequences of these changes for children, families, and society.

43

ACKNOWLEDGMENTS An earlier version of this paper was presented at the annual meeting of the Population Association of America, March, 1989, Baltimore, Md. I wish to thank Daniel T. Lichter, Joseph Lee Rodgers, and two anonymous reviewers for helpful comments on earlier drafts of this paper.


REFERENCES ADAMS, B. N. 1972. Birth order: A critical review. Sociometry 35:411-439. BENGTSON, V. L., and D. DANNEFER. 1987. Fam-

ilies, work and aging: Implications of disordered cohort flow for the twenty-first century. In R. A. Ward and S. S. Tobin (eds.), Health in aging: Sociological issues and policy directions. Springer, New York. BLAKE, J. 1981. Family size and the quality of children. Demography 18:421-442. . 1985. Number of siblings and educational mobility. Amer. Sociol. Rev. 50:8493. . 1989. Family size and achievement. Univ. of California Press, Berkeley.

HASTINGS, D. W., and W. W. ROBINSON. 1975.

Open and closed birth intervals for oncemarried spouse-present white women. Demography 12:455-466. HERNANDEZ, D. 1986. Children and their extended families since World War II. Paper presented at the annual meeting of the Population Association of America in San Francisco, April. HEUSER, R. 1967. Multiple births, 1964. Vital health statistics, data from the National Vital Statistics System, Series 21, No. 14. National Center for Health Statistics, Rockville, Md. HOFFERTH, S. L., and D. A. PHILLIPS. 1987.

Child care in the United States, 1970 to 1995, J. Marr. Fam. 49:559-571. JONES, E. F. 1981. The impact of women's employment on marital fertility in the U.S., 1970-75. Pop. Stud. 35:161-173. KAMMEYER, K. 1967. Birth order and the feminine sex role among college women. Amer. EDWARDS, J. N., and M. B. BRAUNBURGER. 1973. Sociol. Rev. 31:508-515. Exchange and parent-youth conflict. J. Marr. Fam. 35:101-107. KIDWELL, J. S. 1981. Number of siblings, sibling spacing, sex, and birth order: Their effects on EGGEBEEN, D. J. 1986. Parents and siblings of perceived parent-adolescent relationships. J. America's preschool children: 1940-1980. Marr. Fam. 32:216-226. Ph.d. thesis, Dept. of Sociology. University of North Carolina-Chapel Hill, Chapel Hill, . 1982. The neglected birth order: MiddleNC. borns. J. Marr. Fam. 44:225-235. . 1988. Determinants of maternal employ- LASKO, J. K. 1954. Parent behavior toward first ment for white preschool children: 1960and second children. Gen. Psych. Mono. 1980. J. Marr. Fam. 50:149-159. 49:97-137. . 1991. Family structure and intergenera- LEWIS, M., and V. S. KREITZBERG. 1980. Effects tional exchanges. Paper presented at the anof birth order and spacing on mother-infant nual meeting of the Population Association interactions. Dev. Psych. 15:617-625. of America, Washington, D.C., March. MENKEN, J. 1975. Biometric models of fertility. EDLER, G., and C. E. BOWERMAN. 1963. Family Soc. Forces 54:52-65. structure and child rearing patterns: The ef- NATIONAL CENTER FOR HEALTH STATISTICS. 1988. fect of family size and sex composition. Vital statistics of the United States, 1986, Amer. Sociol. Rev. 28:891-905. Vol. I, Natality, Table 1-6. U.S. Government Printing Office, Washington, D.C. FEATHERMAN, D. L., and R. M. HAUSER. 1978. Opportunity and change. Academic Press, NYE, F. I., J. CARLSON, and G. GARRETT. 1970. New York. Family size, interaction, affect, and stress. J. Marr. Fam. 32:216-226. GROAT, H. T. 1982. Working women and childbearing: United States vital and health sta- PETERSON, E. T., and P. R. KUNZ. 1975. Parentistics. Series 23(9), February. tal control over adolescents according to famHAGESTAD, G. O. 1986. The aging society as a ily size. Adolescence 10:419-427. context for family life. Daedalus 115:119- POTTER, R. G. 1963. Birth intervals: Structure 140. and change. Pop. Stud. 17:155-166. BOSSARD, J. H., and E. S. BOLL. 1960. The soci-

ology of child development. Harper, New York. BRIM, O. G. 1958. Family structure and sex role learning by children: A further analysis of Helen Koch's data. Sociometry 21:1-16.

44

Eggebeen

PRESTON, S. H. 1976. Family sizes of children and family sizes of women. Demography 13:105-114. . 1984. Children and the elderly: Divergent paths for America's dependents. Demography 21:435-457. REISS, I. L. 1967. The social context of premarital sexual permissiveness. Holt, Rinehart, and Winston, New York.


RINDFUSS, R., and J. SWEET. 1977. Postwar fer-

tility trends and differentials in the United States. Academic Press, New York. RODGERS, J. L. 1983. Family configuration and adolescent sexual behavior. Pop. Environ. 6:73-83. RODGERS, J. L., and D. C. ROWE. 1988. Influ-

ence of siblings on adolescent sexual behavior. Develop. Psych. 24:722-728. RODGERS, J. L., and V. D. THOMPSON. 1985/86.

Toward a general framework of family structure: A review of theory based empirical research. Pop. Environ. 8:143-172.

Social Biology

I. L. Reiss (eds.), Contemporary theories about the family (Vol. 1). The Free Press, New York. SPANIER, G., P. A. ROSS, and J. SHOCKEY. 1985.

Marital trajectories of American women: Variations in the life course. J. Marr. Fam. 47:993-1003. Tsui, A. O. 1982. The family formation process among U.S. marriage cohorts. Demography 19:1-28. U.S. BUREAU OF THE CENSUS. 1978. Trends in

childspacing: June, 1975. Series P-20, No. 315, February, Current Population Reports, U.S. Government Printing Office, Washington, D.C. . 1988. Marital status and living arrangements: March 1987. Series P-20, No. 423, March, Current Population Reports, U.S. Government Printing Office, Washington, D.C. WAGNER, M. E., H. I. SCHUBERT, and D. S. P.

young male adolescent's perception of early childbearing behavior: The differential effects of socioeconomic status and family size. Sociometry 39:39-52. SCHOOLER, C. 1972. Birth order effects: Not here, not now! Psychol. Bull. 78:161-175.

SCHUBERT. 1979. Sibling constellation effects on psychosocial development, creativity and health. Adv. Child Dev. Behav. 14:57-148. WESTOFF, C. F. 1978. Some speculations on the future of marriage and fertility. Fam. Plan. Perspec. 10:78- 83. ZAJONC, R. B. 1976. Family configuration and intelligence. Science 192:227-236.

SCHVANEVELDT, J. D., and M. INHINGER. 1979.

ZAJONC, R. B., H. MARKUS, and G. B. MARKUS.

Sibling relationships in the family, p. 453467. In W. R. Burr, R. Hill, F. I. Nye, and

The order puzzle. J. Personal. Soc. Psych. 37:1325-1341.

SCHECK, D. C., and R. EMERICK. 1976. The

Social values and preschool behavioral adjustment: A comparative investigation of Latino and European American preschool children.

Screening for language disorders in preschool children.

Hospital admissions for accidents in preschool children.

Hospital admissions for accidents in preschool children.

Changes in sibling behavior following family intervention.

Change in Misperception of Child's Body Weight among Parents of American Preschool Children.

Behavior management for preschool-aged children.

Vision screening in preschool children.

Automated hearing screening for preschool children.

Vision screening in preschool children.

Managing wheeze in preschool children.

Vision screening in preschool children.

Longitudinal changes in individual symptoms across the preschool years in children with ADHD.

Treatment strategies for ADHD in preschool and school-age children.

Macrolides for Acute Wheezing Episodes in Preschool Children.

Characteristics of Print in Books for Preschool Children.

Nutritional status of preschool children.

Bedtime in Preschool-Aged Children and Risk for Adolescent Obesity.

Secular Changes of Adiposity and Motor Development in Czech Preschool Children: Lifestyle Changes in Fifty-Five Year Retrospective Study.

Binocular iPad treatment for amblyopia in preschool children.

Pregnancy and post-delivery maternal weight changes and overweight in preschool children.

Effect of WIC Food Package Changes on Dietary Intake of Preschool Children in New Mexico.

Health needs of preschool children.

Pharmacological approach to wheezing in preschool children.